[SATLUG] Software RAID suggestions

FIRESTORM_v1 firestorm.v1 at gmail.com
Wed Mar 3 03:41:47 CST 2010

ok, so maybe writing email over RDP over SSH isn't the best idea...

- the raid card IS in JBOD mode now for this rebuild

- John, I would be very interested in your layout information that
bounced.  Please send it to me if you still have it written up.

Thank you.


On Wed, Mar 3, 2010 at 3:29 AM, FIRESTORM_v1 <firestorm.v1 at gmail.com> wrote:
> Hey all, well I _finally_ gave up on trying to use the RAID card as a
> RAID card and figured I can beat the snot out of software raid before
> I let the server go live to check on data errors, etc.
> I noticed something odd about this setup and after a while, I did end
> up killing this build.  I did some research and noticed that if I used
> --force (to force mdadm to use only the drives I specified) mdstat
> showed [UUUU] and appears to be rebuilding properly.  An additional
> change I made is I went in and created partitions and made them all
> type fd (linux raid auto) and went into the RAID card's BIOS and made
> sure that the disks weren't in JBOD mode.  Although Linux saw the four
> drives, it really freaked out on the partition types but fdisk led me
> to believe the drives were ready to go.
> Now, the only concern I have is that when the server reboots, will
> mdadm mount the RAID using all four drives like I specified or will it
> break the array at boot requiring me to rebuild it again manually?
> When I did the reinstall after making sure that the card was set for
> JBOD, the installer saw the four drives but proceeded to mark the
> fourth offline. [UUU_]
> When I used mdadm --stop /dev/md0 and then respecified my mdadm create
> command with --force it did not generate any errors and is creating
> the array as I designated with all drives marked online.  I guess now,
> it's the waiting game as I wait an almost 36 hours for the array to
> build.
> Since we're on the subject, can anyone recommend a PCI 4 port SATA
> card that doesn't have RAID on it and that works well with Linux?  I
> found a $30 card that uses the silicon labs chipset, but the most
> recent data I could find on it regarding linux support was from over
> two years ago and that it wasn't very good.
> Also, (and I'm sure this is a stupid as heck question) since I'm using
> software RAID, would a faster proc mean better throughput?  Right now
> the "server" is a P-II 300MHz with 512MB RAM and was wondering if I
> will need to dig a better machine out of storage (think I have a P-III
> 850 somewhere using 256MB RDRAM) but getting it out of storage is
> going to take an act of god and an act of congress.
> This isn't going to be a "high load" server, just a fileserver for
> XBMC and windows files.
> Thank you all for your suggestions, it's definitely a learning experience.
> On Fri, Feb 26, 2010 at 4:51 PM, Jeremy Mann <jeremymann at gmail.com> wrote:
>> On Fri, Feb 26, 2010 at 1:39 PM, FIRESTORM_v1 <firestorm.v1 at gmail.com> wrote:
>>> My only concern is that the third line [UUU_] looks a little odd.   I
>>> expected [UUUU] since I have 4 SATA drives and I specified their
>>> devices on the commandline:
>>> #  mdadm --create /dev/md0 --level=5 --raid-devices=4 /dev/sdb
>>> /dev/sdc /dev/sdd /dev/sde
>>> I don't recall anyone specifying what the [UUU_] but can someone tell
>>> me if I'm chasing ghosts or letting myself get too nervous?
>> Matt, "_" means one drive is offline, but that could be the one that
>> is rebuilding. Once its finished check it again with :
>> cat /proc/mdstat
>> If it still shows one drive as "_", you got problems and your RAID
>> array is running in degraded mode.
>> --
>> Jeremy Mann
>> jeremy at biochem.uthscsa.edu
>> University of Texas Health Science Center
>> Bioinformatics Core Facility
>> http://www.bioinformatics.uthscsa.edu
>> Phone: (210) 567-2672
>> --
>> _______________________________________________
>> SATLUG mailing list
>> SATLUG at satlug.org
>> http://alamo.satlug.org/mailman/listinfo/satlug to manage/unsubscribe
>> Powered by Rackspace (www.rackspace.com)

More information about the SATLUG mailing list